Asbestos-free Hybrid Smart Tribological Design Of The Fiber Friction Material

In this paper the optimal design of non-asbestos hybrid fiber friction materials is dealt with the artificial intelligence technique. After the training-sets have been determined by the method of uniform design, a neural network is established from friction coefficient to friction material composition. Then the genetic algorithm is chosen to search the optimal material formulation. Utilizing the neural network toolbox and the programming packet of genetic algorithm in Matlab, the cooperating optimization of neural network and genetic algorithm is easily programmed. The results show that the model is able to predict the mapping relation of friction coefficient and material composition. Also they suggest the genetic algorithm has searched the optimal formulation which has good fade properties. Compared with the national standard, the performance index of this optimal formulation is satisfied with the need of automobile brakes. And its composition is similar to the results of orthogonal design. So the cooperating optimization of neural network and genetic algorithm is capable of replacing the tradition design methods and realizes the intelligence design of friction materials.